ai-ml-attacks
This skill should be used when the user asks about "AI security", "ML pipeline attacks", "prompt injection", "model deserialization", "unsafe model loading",…
This skill should be used when the user asks about "API security", "OWASP API Top 10", "BOLA", "broken object level authorization", "API authentication", "mass assignment", "GraphQL security", "gRPC security", "rate limiting", "API abuse", "REST API vulnerabilities", or needs to
$ npx -y skills add allsmog/vuln-scout --skill owasp-api-top10 --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/owasp-api-top10Context preview
The summary Claude sees to decide when to auto-load this skill.
This skill should be used when the user asks about "API security", "OWASP API Top 10", "BOLA", "broken object level authorization", "API authentication", "mass assignment", "GraphQL security", "gRPC security", "rate limiting", "API abuse", "REST API vulnerabilities", or needs to
name: OWASP API Security Top 10 description: This skill should be used when the user asks about "API security", "OWASP API Top 10", "BOLA", "broken object level authorization", "API authentication", "mass assignment", "GraphQL security", "gRPC security", "rate limiting", "API abuse", "REST API vulnerabilities", or needs to identify API-specific security issues during whitebox security review. version: 1.0.0
Identify API-specific security vulnerabilities based on the OWASP API Security Top 10 (2023 edition), including REST, GraphQL, and gRPC patterns. APIs are the primary attack surface for modern applications, and these vulnerability classes represent the most critical and commonly exploited API security risks.
Activate this skill when:
**Description**: API endpoints that accept object IDs and do not verify that the authenticated user has permission to access the referenced object.
**Pattern**: ID parameters used directly in database queries without ownership checks.
**Vulnerable Code Examples**:
Express.js:
// VULNERABLE: No ownership check
app.get('/api/orders/:id', auth, async (req, res) => {
const order = await Order.findById(req.params.id);
res.json(order); // Any authenticated user can access any order
});Flask:
# VULNERABLE: No ownership check
@app.route('/api/orders/<int:order_id>')
@login_required
def get_order(order_id):
order = Order.query.get_or_404(order_id)
return jsonify(order.to_dict()) # Missing: order.user_id == current_user.idSpring:
// VULNERABLE: No ownership check
@GetMapping("/api/orders/{id}")
public Order getOrder(@PathVariable Long id) {
return orderRepository.findById(id)
.orElseThrow(); // Any authenticated user can access any order
}Go (net/http):
// VULNERABLE: No ownership check
func getOrder(w http.ResponseWriter, r *http.Request) {
id := mux.Vars(r)["id"]
order, _ := db.GetOrder(id) // Missing: order.UserID == authenticatedUser.ID
json.NewEncoder(w).Encode(order)
}**Detection Patterns**:
# Express: Route params used directly in DB queries
grep -rniE "req\.params\.\w+" --include="*.js" --include="*.ts" -A 5 | grep -iE "(findById|findOne|findByPk|where)"
# Flask: Route parameters in DB queries without ownership check
grep -rniE "def\s+\w+\(.*_id\)" --include="*.py" -A 10 | grep -iE "(query\.get|filter_by|find_one)"
# Spring: PathVariable in repository calls
grep -rniE "@PathVariable" --include="*.java" -A 5 | grep -iE "(findById|getOne|findOne)"
# Go: URL parameters in DB calls
grep -rniE "(mux\.Vars|chi\.URLParam|r\.URL\.Query)" --include="*.go" -A 5 | grep -iE "(GetOrder|FindBy|QueryRow)"
# Generic: Look for endpoints with ID params lacking authorization
grep -rniE "(/:id|/<int:|/{id}|\{id\})" --include="*.js" --include="*.ts" --include="*.py" --include="*.java" --include="*.go"**Fix Pattern**: Always verify that the authenticated user owns or has permission to access the requested resource:
order = Order.query.get_or_404(order_id)
if order.user_id != current_user.id:
abort(403)**Description**: Weak or missing authentication mechanisms on API endpoints.
**Vulnerable Patterns**:
| Pattern | Risk | Detection | |---------|------|-----------| | API key in URL query string | HIGH | `grep -rniE "api[_-]?key=.*[?&]" --include="*.js" --include="*.py"` | | No rate limiting on auth endpoints | HIGH | Check `/login`, `/auth`, `/token` for rate limit middleware | | JWT with `none` algorithm | CRITICAL | `grep -rniE "algorithm.*none\|alg.*none" --include="*.py" --include="*.js"` | | Missing token expiration | HIGH | `grep -rniE "expiresIn\|exp" --include="*.js" --include="*.py"` (check if absent) | | Hardcoded JWT secret | CRITICAL | `grep -rniE "(jwt_secret\|JWT_SECRET\|secret_key)\s*=\s*[\"']" --include="*.py" --include="*.js"` | | Password in API response | HIGH | Check serialization excludes password fields | | No authentication on sensitive endpoints | CRITICAL | Routes without `auth`, `authenticate`, `login_required` middleware |
**Per-Framework Detection**:
# Express: Routes without auth middleware
grep -rniE "app\.(get|post|put|delete|patch)\s*\(" --include="*.js" --include="*.ts" | grep -viE "(auth|protect|verify|middleware|passport)"
# Flask: Routes without login_required
grep -rniE "@app\.route" --include="*.py" -A 2 | grep -viE "(login_required|jwt_required|auth)"
# Spring: Endpoints without @PreAuthorize or security config
grep -rniE "@(Get|Post|Put|Delete)Mapping" --include="*.java" -B 3 | grep -viE "(PreAuthorize|Secured|RolesAllowed)"
# Go: Handlers without auth middleware in chain
grep -rniE "HandleFunc|Handle\(" --include="*.go" | grep -viE "(auth|middleware|protect)"**Description**: API exposes object properties that the user should not be able to read (excessive data exposure) or write (mass assignment).
**Excessive Data Exposure Patterns**:
# Returning entire ORM objects without field filtering grep -rniE "\.to_dict\(\)|\.toJSON\(\)|jsonify\(\w+\)|res\.json\(\w+\)" --include="*.py" --include="*.js" # Spring: Returning entity directly grep -rniE "return\s+\w+Repository\.find" --include="*.java" # Missing field selection in serialization grep -rniE "JSON\.stringify\(user\)|json\.dumps\(\w+\.__dict__\)" --include="*.js" --include="*.py"
**Mass Assignment Patterns**:
# Express: Spreading request body into model grep -rniE "\.c
AI-powered whitebox penetration testing plugin for Claude Code. 9 languages, 22 skills, 7 autonomous agents. STRIDE threat modeling, OWASP 2025 coverage, polyglot monorepo support.
Repo: allsmog/vuln-scout
This skill should be used when the user asks about "AI security", "ML pipeline attacks", "prompt injection", "model deserialization", "unsafe model loading",…
This skill should be used when the user asks about "business logic", "workflow vulnerability", "trust boundary", "state machine", "authorization bypass",…
This skill should be used when the user asks about "cache poisoning", "web cache deception", "CDN cache", "proxy cache", "nginx cache", "varnish", "cache key…
This skill should be used when the user asks about "cloud security", "AWS security", "GCP security", "Azure security", "Kubernetes security", "IMDS", "instance…
This skill should be used when the user asks about "compliance mapping", "PCI-DSS", "HIPAA", "SOC 2", "NIST CSF", "regulatory requirements", "compliance…
This skill should be used when the user asks about "Code Property Graph", "CPG analysis", "Joern queries", "CPGQL", "data flow verification", "taint tracking…